An multi-period optimisation tool to promote solar thermal into district heating networks
Création d’un outil d’optimisation multi-période pour améliorer l’intégration du solaire dans les réseaux de chaleur
Résumé
Regarding the global warming caused by human activity and the urgent need to reduce greenhouse gas emissions, solutions to overcome the ongoing fossil fuels phase-out have to be found. District heating networks have already proved their relevance to mutualise and then enhance renewable resources that are inaccessible or difficult to access in urban areas, especially solar thermal energy. However they are still underdeveloped in the world today, especially in France. The use of such systems, which require important investments, is still low. One of the impediments to this development is the lack of tools and knowledge to optimize the integration of solar energy in multi-source district heating networks. The objective of this thesis, financed by ADEME and Region Nouvelle-Aquitaine in partnership with a consortium of companies, is to optimize the integration of solar thermal energy via an "open-access" tool that can be used to size and operate the heat production of such networks.The different production elements of a solar district heating network were modelled: a solar thermal field, two boilers (gas and wood) and a thermal storage. A dynamic multi-period approach was chosen in order to transform, by discretization, an algebro-differential model into an algebraic one. Likewise, a method for creating characteristic days was implemented in order to represent a year with a minimum number of days. The algebraic equations of the resulting model are non linear. All the variables are continuous. The chosen objective function is economic: the project partners provided the input data for an accurate economic model using non linear equations. Finally the formulated problem is a NLP (Non Linear Programming) problem. As non-linear optimization guarantees neither systematic convergence nor a global optimum, a solution strategy was developed to guarantee a confidence optimum.A first application case optimizing the sizing and operation of a new network using inter-seasonal storage was studied. The first results showed a certain robustness of the resolution strategy and converged towards a sizing that valorises the solar energy up to 50% of the total energy supplied over the lifetime of the installation, which was set at 20 years. Different storage connections were studied to conclude on the relevance of connecting the storage at the output of the solar field. Sensitivity analyses on the minimum renewable rate and the lifetime of the installation were carried out, highlighting that it is possible to greatly increase the renewable rate with moderate additional investments.In some cases (very dense urban area for example) the implementation of inter-seasonal storage and a large solar field may be impossible. Then, a second application case, an optimization of a new network using daily storage, was studied. This optimization led to a result where solar energy covers 12 % of the total energy supplied with a smaller solar field. These applications show that solar thermal can cover a large part of demand (>50%) when the use of inter-seasonal storage and a large solar field is possible, but that it also presents an interest in covering a smaller but nonetheless not insignificant part of the demand with daily storage.
L’objectif de ces travaux est de développer une méthodologie pour l'optimisation dynamique d'un système intégré incluant la production, le stockage et la distribution de chaleur. Un tel système est soumis à de nombreuses perturbations à différentes échelles de temps. L'objectif est donc de développer un outil d'aide à la conception d'un système dynamique complexe. Cette thèse vise dans un premier temps à développer un modèle dynamique du système global. Ensuite, le problème d'optimisation dynamique avec variables mixtes sera formulé puis résolu.
Origine : Version validée par le jury (STAR)